问题
Following the answer from a previous question I want to get rid of some NULL values when I do not have a number for some cases
I have the following query
CREATE TABLE [MYTABLE]
(
[A] int
, [B] int
, [VAL] float
);
INSERT INTO [MYTABLE] ([A], [B], [VAL]) VALUES
(1, 1, -218.46),
(1, 2, -2.1846000000000e+02),
(2, 1, 6.4000000000000e+01),
(2, 2, -218.46),
(2, 3, 6.4000000000000e+01),
(3, 2, -2.1846000000000e+02),
(3, 3, -218.46),
(4, 3, 6.4000000000000e+01),
(4, 4, -218.46),
(4, 5, 6.4000000000000e+01),
(5, 4, 6.4000000000000e+01),
(5, 5, -218.46),
(5, 6, 6.4000000000000e+01),
(6, 5, 6.4000000000000e+01),
(6, 6, -218.46),
(6, 7, 6.4000000000000e+01),
(7, 6, 6.4000000000000e+01),
(7, 7, -218.46),
(7, 8, 6.4000000000000e+01),
(8, 7, 6.4000000000000e+01),
(8, 8, -218.46),
(8, 9, 6.4000000000000e+01),
(9, 8, 6.4000000000000e+01),
(9, 9, -218.46),
(9, 10, 6.4000000000000e+01),
(10, 9, 6.4000000000000e+01),
(10, 10, -218.46);
SELECT [1],[2],[3],[4],[5],[6],[7],[8],[9],[10]
FROM
( SELECT x = a
, y = b
, v = (CASE WHEN IsNumeric(val)=0 THEN 0 ELSE val END)
FROM MYTABLE
) AS T
PIVOT
(
min( v)
FOR y IN ([1],[2],[3],[4],[5],[6],[7],[8],[9],[10])
) AS pvt
order by x
I am getting the following result:
1 2 3 4 5 6 7 8 9 10
------- ------- ------- ------- ------- ------- ------- ------- ------- -------
-218.46 -218.46 NULL NULL NULL NULL NULL NULL NULL NULL
64 -218.46 64 NULL NULL NULL NULL NULL NULL NULL
NULL -218.46 -218.46 NULL NULL NULL NULL NULL NULL NULL
NULL NULL 64 -218.46 64 NULL NULL NULL NULL NULL
NULL NULL NULL 64 -218.46 64 NULL NULL NULL NULL
NULL NULL NULL NULL 64 -218.46 64 NULL NULL NULL
NULL NULL NULL NULL NULL 64 -218.46 64 NULL NULL
NULL NULL NULL NULL NULL NULL 64 -218.46 64 NULL
NULL NULL NULL NULL NULL NULL NULL 64 -218.46 64
NULL NULL NULL NULL NULL NULL NULL NULL 64 -218.46
but I need the following result
1 2 3 4 5 6 7 8 9 10
------- ------- ------- ------- ------- ------- ------- ------- ------- -------
-218.46 -218.46 NULL NULL 0 0 0 0 0 0
64 -218.46 64 0 0 0 0 0 0 0
0 -218.46 -218.46 0 0 0 0 0 0 0
0 0 64 -218.46 64 0 0 0 0 0
0 0 0 64 -218.46 64 0 0 0 0
0 0 0 0 64 -218.46 64 0 0 0
0 0 0 0 0 64 -218.46 64 0 0
0 0 0 0 0 0 64 -218.46 64 0
0 0 0 0 0 0 0 64 -218.46 64
0 0 0 0 0 0 0 0 64 -218.46
How do I replace all NULL values with 0's?
EDIT
Following @Siva answer
CREATE TABLE [dbo].[Matrix]
(
[A] INT
, [B] INT
, [VAL] FLOAT
);
INSERT INTO [dbo].[Matrix] ([A], [B], [VAL])
VALUES
(1, 1, -218.46)
, (1, 2, -2.1846000000000e+02)
, (2, 1, 6.4000000000000e+01)
, (2, 2, -218.46)
, (2, 3, 6.4000000000000e+01)
, (3, 2, -2.1846000000000e+02)
, (3, 3, -218.46)
, (4, 3, 6.4000000000000e+01)
, (4, 4, -218.46)
, (4, 5, 6.4000000000000e+01)
, (5, 4, 6.4000000000000e+01)
, (5, 5, -218.46)
, (5, 6, 6.4000000000000e+01)
, (6, 5, 6.4000000000000e+01)
, (6, 6, -218.46)
, (6, 7, 6.4000000000000e+01)
, (7, 6, 6.4000000000000e+01)
, (7, 7, -218.46)
, (7, 8, 6.4000000000000e+01)
, (8, 7, 6.4000000000000e+01)
, (8, 8, -218.46)
, (8, 9, 6.4000000000000e+01)
, (9, 8, 6.4000000000000e+01)
, (9, 9, -218.46)
, (9, 10, 6.4000000000000e+01)
, (10, 9, 6.4000000000000e+01)
, (10, 10, -218.46);
SELECT COALESCE([1], 0) [1]
, COALESCE([2], 0) [2]
, COALESCE([3], 0) [3]
, COALESCE([4], 0) [4]
, COALESCE([5], 0) [5]
, COALESCE([6], 0) [6]
, COALESCE([7], 0) [7]
, COALESCE([8], 0) [8]
, COALESCE([9], 0) [9]
, COALESCE([10], 0) [10]
FROM
(
SELECT x = a
, y = b
, v = val
FROM [dbo].[Matrix]
) AS dataToPivot
PIVOT
(
MIN(v)
FOR y IN ([1], [2], [3], [4], [5], [6], [7], [8], [9], [10])
) AS pivoted
ORDER BY x;
What woul be the best way to do this in a dynamic for, I was thinking to do FOR THE FIRST PART...:
DECLARE @sql nvarchar(max);
DECLARE @Flag INT;
DECLARE @Size INT;
SET @Flag = 1
SET @Size = 10 ;
SET @sql = 'SELECT '+ CHAR(13);
WHILE (@Flag <= @Size) BEGIN
SET @sql = @sql + ' COALESCE([' + cast( @Flag as varchar(10)) + '], 0) ['+ cast( @Flag as varchar(10))+'],'+ CHAR(13) ;
SET @Flag = @Flag + 1;
END
SET @sql = LEFT(@sql, LEN(@sql) - 2)
SET @sql = @sql + '
FROM
(
SELECT x = a
, y = b
, v = val
FROM [dbo].[Matrix]
) AS dataToPivot
PIVOT
(
MIN(v)
FOR y IN (';
SET @Flag = 1
WHILE (@Flag <= @Size) BEGIN
SET @sql = @sql + ' [' + cast( @Flag as varchar(10)) + '],';
SET @Flag = @Flag + 1;
END
SET @sql = LEFT(@sql, LEN(@sql) - 1);
SET @sql = @sql +
')
) AS pivoted
ORDER BY x;'
print @sql
exec(@sql);
回答1:
You can use COALESCE to achieve the result.
SQL Fiddle
CREATE TABLE [dbo].[Matrix]
(
[A] INT
, [B] INT
, [VAL] FLOAT
);
INSERT INTO [dbo].[Matrix] ([A], [B], [VAL])
VALUES
(1, 1, -218.46)
, (1, 2, -2.1846000000000e+02)
, (2, 1, 6.4000000000000e+01)
, (2, 2, -218.46)
, (2, 3, 6.4000000000000e+01)
, (3, 2, -2.1846000000000e+02)
, (3, 3, -218.46)
, (4, 3, 6.4000000000000e+01)
, (4, 4, -218.46)
, (4, 5, 6.4000000000000e+01)
, (5, 4, 6.4000000000000e+01)
, (5, 5, -218.46)
, (5, 6, 6.4000000000000e+01)
, (6, 5, 6.4000000000000e+01)
, (6, 6, -218.46)
, (6, 7, 6.4000000000000e+01)
, (7, 6, 6.4000000000000e+01)
, (7, 7, -218.46)
, (7, 8, 6.4000000000000e+01)
, (8, 7, 6.4000000000000e+01)
, (8, 8, -218.46)
, (8, 9, 6.4000000000000e+01)
, (9, 8, 6.4000000000000e+01)
, (9, 9, -218.46)
, (9, 10, 6.4000000000000e+01)
, (10, 9, 6.4000000000000e+01)
, (10, 10, -218.46);
SELECT COALESCE([1], 0) [1]
, COALESCE([2], 0) [2]
, COALESCE([3], 0) [3]
, COALESCE([4], 0) [4]
, COALESCE([5], 0) [5]
, COALESCE([6], 0) [6]
, COALESCE([7], 0) [7]
, COALESCE([8], 0) [8]
, COALESCE([9], 0) [9]
, COALESCE([10], 0) [10]
FROM
(
SELECT x = a
, y = b
, v = val
FROM [dbo].[Matrix]
) AS dataToPivot
PIVOT
(
MIN(v)
FOR y IN ([1], [2], [3], [4], [5], [6], [7], [8], [9], [10])
) AS pivoted
ORDER BY x;
回答2:
How about IsNull?
SELECT
isnull([1], 0) as [1],
isnull([2], 0) as [2],
isnull([3], 0) as [3],
isnull([4], 0) as [4],
isnull([5], 0) as [5],
isnull([6], 0) as [6],
isnull([7], 0) as [7],
isnull([8], 0) as [8],
isnull([9], 0) as [9],
isnull([10], 0) as [10]
FROM
(SELECT x=a, y=b, v= CASE WHEN IsNumeric(val)=0 THEN 0 ELSE IsNull(val, 0) END FROM MYTABLE)
as T
PIVOT
(
min(v)
FOR y IN ([1],[2],[3],[4],[5],[6],[7],[8],[9],[10])
) AS pvt
ORDER BY x
回答3:
I used the dynamic pivot technique in this SO question Simple way to transpose columns and rows in Sql?
To put NULLs in my pivot report like below. Does get more messy though.
CREATE TABLE [MYTABLE]
(
[A] int
, [B] int
, [VAL] float
);
INSERT INTO [MYTABLE] ([A], [B], [VAL]) VALUES
(1, 1, -218.46),
(1, 2, -2.1846000000000e+02),
(2, 1, 6.4000000000000e+01),
(2, 2, -218.46),
(2, 3, 6.4000000000000e+01),
(3, 2, -2.1846000000000e+02),
(3, 3, -218.46),
(4, 3, 6.4000000000000e+01),
(4, 4, -218.46),
(4, 5, 6.4000000000000e+01),
(5, 4, 6.4000000000000e+01),
(5, 5, -218.46),
(5, 6, 6.4000000000000e+01),
(6, 5, 6.4000000000000e+01),
(6, 6, -218.46),
(6, 7, 6.4000000000000e+01),
(7, 6, 6.4000000000000e+01),
(7, 7, -218.46),
(7, 8, 6.4000000000000e+01),
(8, 7, 6.4000000000000e+01),
(8, 8, -218.46),
(8, 9, 6.4000000000000e+01),
(9, 8, 6.4000000000000e+01),
(9, 9, -218.46),
(9, 10, 6.4000000000000e+01),
(10, 9, 6.4000000000000e+01),
(10, 10, -218.46);
DECLARE @Sql nvarchar(max), @PvtCols nvarchar(max), @PvtSelect nvarchar(max)
SELECT @PvtSelect = Stuff(
(
SELECT ',IsNull(' + Quotename(B) + ',0) as ' + Quotename(B)
FROM MYTABLE mt
GROUP BY B
ORDER BY B
FOR XML PATH(''), TYPE
).value('.', 'NVARCHAR(MAX)')
,1,1,'')
SELECT @PvtCols = Stuff(
(
SELECT DISTINCT ',' + Quotename(B)
FROM MYTABLE mt
FOR XML PATH(''), TYPE
).value('.', 'NVARCHAR(MAX)')
,1,1,'')
SET @Sql =
'SELECT ' + @PvtSelect + '
FROM
( SELECT x = a
, y = b
, v = (CASE WHEN IsNumeric(val)=0 THEN 0 ELSE val END)
FROM MYTABLE
) AS T
PIVOT
(
min( v)
FOR y IN (' + @PvtCols + ')
) AS pvt
order by x'
EXEC(@SQL)
来源：https://stackoverflow.com/questions/10325538/how-to-replace-the-null-values-in-a-pivot-output-with-zeroes